Papers

Estimating spatially heterogeneous elastic properties from low-resolution displacement measurements is a severely ill-posed inverse elasticity problem because low resolution obscures spatial details needed to distinguish heterogeneous property variations, and small measurement perturbations or fitting errors are amplified through inverse estimation. Existing inverse methods often rely on high-fidelity observations and manually prespecified loss weights, limiting their adaptability and making them sensitive to noise and resolution degradation. We propose a Probabilistic Inverse Elasticity Physics-Informed Neural Network (PIE-PINN) framework for robust estimation of Young's modulus and Poisson's ratio from noisy, low-resolution displacement data. PIE-PINN models displacement observation, strain-discrepancy, and equilibrium residuals using Laplace distributions within a unified probabilistic model. To improve robustness, the framework combines a B-spline-guided displacement network with a hierarchical half-Cauchy model for displacement residual scales. The B-spline provides a smooth global representation of the displacement field, while the neural network correction captures local variations. The hierarchical scale model adaptively downweights severe displacement fitting errors, enabling more robust recovery of the latent mean displacement field. An alternating maximum-likelihood training strategy updates the mean through weighted residual minimization and updates the scales to adjust the loss weights. Systematic case studies across varying noise levels and observation resolutions demonstrate the robustness of PIE-PINN.

Machine-learned predictions can speed up offline NP-hard optimization, but asking a predictor what to do amounts to asking it to solve the problem, and committing an unchecked prediction forfeits every worst-case guarantee. CASP (Certificate-Augmented Solution Pruning) instead asks which parts of the search space may be ignored, and accepts each answer only after a sound polynomial-time verifier has checked it, so correctness never depends on prediction quality. We develop the learning theory of this design. The verifier makes the induced loss class uniformly bounded, so certificate parameters are learnable from $\tilde O(\varepsilon^{-2}\log K)$ samples ($K$ the maximum instance size), whereas the unverified commitment class admits no distribution-free rate and, under cost spread $R$, none below $\Omega(R/\varepsilon^2)$. Filtering noisy predictions by verifiable confidence dominates the standard min-combiner, with a margin we compute in closed form, and the prediction stays useful even given the LP, because it breaks ties on degenerate optimal faces, where every symmetric LP policy, meaning one whose commitments depend on the instance only through the verifiable confidence values, provably stalls. Experiments on five problems test the theory's quantitative predictions. With trained predictors, unverified pruning loses up to $26%$ of the optimum under distribution shift, while the verified deployment of the same predictions loses nothing.

Rich internal representations of musical structure are essential for music understanding tasks such as machine-assisted music co-writing, yet self-supervised approaches for symbolic music representation remain underexplored, particularly those that encode the hierarchical multiscale nature of musical structures. We present MIDI-RAE-JEPA, combining a pitch- and time-shift equivariance objective with LeJEPA and a Swin Transformer V2 encoder to learn such hierarchical representations of symbolic music encoded as piano roll images. The time-shift equivariance objective encourages the model to internalize temporal musical relationships. The encoder is trained purely on self-supervised objectives – including a masked embedding predictor (MEP) – with collapse prevented via SIGReg. A separate decoder trained on the frozen encoder embeddings achieves reconstruction F1 of 0.995, and a flow matching generative model conditioned on those embeddings produces generations that closely match the pitch register and rhythmic density of the conditioning excerpt, while mismatched conditioning yields unrelated but musically plausible output. Learned representations outperform a Haar scattering transform baseline on a downstream emotion classification task, and embedding distances increase monotonically with pitch and time shift magnitude, confirming measurable equivariance. These results suggest that equivariance-based SSL objectives, combined with sufficient fine-level encoder capacity, provide a viable path toward semantically rich, generatively useful representations of symbolic music.

Muon-style optimizers apply a polar map to matrix momentum, but their updates also depend on the representation of each parameter block before orthogonalization. We study this representation choice as a form of optimizer geometry and introduce {\method}, a family of Muon-style optimizers that shares the same momentum rule and Newton–Schulz backend across native, nearest-square, skinny, and vector representations. Each Frobenius-isometric representation induces a distinct polar steepest-descent geometry, in which the shorter matrix dimension determines the number of supported singular channels, the pullback scaling, and the constants in stochastic nonconvex convergence bounds. In a teacher–student model, curvature collapse and an isotropic Marchenko–Pastur spectral profile connect early-stage dissipation to the represented nuclear-to-squared-Frobenius norm ratio. Pretraining experiments on LLaMA2-130M and LLaMA2-600M, together with fixed-momentum diagnostics, show that balanced non-native representations can match the performance of the native representation, whereas reducing the shorter dimension weakens the scaling and singular-channel support, leading to behavior that increasingly resembles normalized momentum.

Hyper-Connections (HC) expand the residual stream of Transformers into $N$ parallel streams, providing a form of memory scaling beyond model width and depth. Manifold-Constrained HC (mHC) stabilizes this formulation at scale. The large gains from $N{=}1$ to $N{=}4$ suggest residual-stream expansion as a promising scaling axis. However, existing HC-family methods typically stop at $N{=}4$. Our experiments reveal why: scaling mHC beyond this point yields diminishing performance gains and rapidly increasing training cost. We attribute this limitation to two bottlenecks: insufficient write-back information for an expanding number of streams and residual-mixing generation whose cost scales cubically with $N$. To address both bottlenecks, we propose xHC (Expanded Hyper-Connections), the first HC-family method to achieve meaningful expansion beyond $N{=}4$. xHC combines temporal feature augmentation for richer write-back with a sparse residual-stream architecture that updates only $k=4$ of the $N=16$ streams while retaining dense access to the full residual state. Across 18B and 28B MoE models, xHC delivers strong and consistent downstream improvements. On an 18B MoE model, xHC improves the average downstream score by 4.0 points over mHC, while adding only modest training FLOPs over the vanilla baseline. Scaling-law experiments show that the vanilla and mHC require $1.50\times$ and $1.19\times$ the compute of xHC, respectively, to reach the same loss. Practical large-$N$ training also requires controlling memory traffic from the expanded residual state. We therefore introduce xHC-Flash, which reduces the per-sublayer memory traffic from $73.5C$ to $40C$, comparable to the $34C$ required by mHC at $N{=}4$, while retaining the gains of full xHC. Together, xHC and xHC-Flash make large-$N$ residual-stream expansion effective and practical for LLM pre-training.

Stein variational gradient descent (SVGD) transports interacting particles toward a target distribution through deterministic kernelized dynamics. Singular Riesz kernels are attractive because they can provide quantitative population-level convergence, but at the finite-particle level the corresponding Stein energy has infinite self-interaction. We study periodic Riesz SVGD with self-interaction removed and prove a many-particle, long-time sampling theorem. Throughout the range in which the singular Stein energy is locally integrable, under a uniform bound on the initial relative entropy per particle, the time-averaged empirical-measure law converges weakly to the point mass \(\delta_\pi\) at the target as the particle number and any diverging averaging horizon tend to infinity. We also show that the empirical-measure laws induced by invariant particle laws of finite relative entropy converge weakly to \(\delta_\pi\), without a uniform entropy bound. Below the logarithmic singularity threshold, we obtain an explicit algebraic finite-particle error bound. These results extend the joint-entropy approach for smooth-kernel SVGD to singular interactions.

We formulate reinforcement learning (RL) in continuous time with discrete state spaces and possibly arbitrary action spaces via a stochastic control approach, where the state dynamics are modeled as a controlled continuous-time Markov chain (CTMC). We consider policy optimization problems and derive corresponding policy gradient methods, leading to continuous-time variants of proximal policy optimization (PPO) and group relative policy optimization (GRPO). As a primary application, we develop a continuous-time RL framework for fine-tuning score-based discrete diffusion models, which enables reward-driven optimization without requiring differentiability on the reward signals. In contrast to the existing GRPO-based approaches that only rely on terminal rewards, our formulation allows intermediate reward or advantage signals to be incorporated throughout the denoising trajectory. Importantly, when specialized to masked diffusion models (MDMs), our framework encompasses a rich class of policy parameterizations over the vocabulary simplex with analytically tractable probability ratios, providing a unified perspective on exploration and policy optimization in MDMs. For diffusion large language models (dLLMs), we further propose trajectory subsampling techniques to efficiently estimate computationally prohibitive trajectory likelihoods, reducing the computational cost of computing per-position probability ratios. We showcase the effectiveness of our methods on both low-dimensional entropy-regularized optimization problems and RL post-training of dLLMs on reasoning and coding tasks.

Interpreting optimizers as gradient-flow discretizations has motivated applying higher-order Runge-Kutta (RK) integrators to neural networks. We build a representative Adam variant (Bogacki-Shampine 3(2) RK pair, FSAL reuse, local-error step control) and evaluate it under a strict compute-matched protocol giving every method the same gradient-evaluation budget - an accounting this literature rarely enforces. Under it the RK variant loses to plain Adam on training loss in both minibatch and full-batch (RK's best-case) training. Instrumenting it shows the "adaptivity" is illusory: normalized error stays far below tolerance, the step size pins at its growth cap from step one (98-100 percent of steps), and no rtol x hmax x h0 setting makes it act; tolerances spanning 100x give bit-identical trajectories. The method is exactly fixed-step Adam with an averaged gradient at 3-4x cost. Repairing it (true reject branch; error on the applied map) reverses the full-batch result - about 40x lower training loss than tuned Adam - and a fixed-step control isolates adaptivity (an emergent warmup-and-growth schedule) as the mechanism. But the gain is fragile to the initial step size and does not reach test accuracy. A pre-registered follow-up rules out the obvious explanations: deeper minimization does not overfit, and an explicit temperature knob only hurts - leaving a trajectory effect, the controller selecting a minimum generalizing 1.3-3.4 points below first-order descent at equal depth. An n=10 study confirms one secondary effect: gradient averaging is a genuine implicit regularizer, beating lr-matched Adam and AdamW on 10/10 seeds - yet RMSprop and NAdam match or beat it at a third the per-step cost. Higher-order adaptive integration buys deeper deterministic minimization and a small regularization effect, but nothing a cheaper, well-tuned first-order baseline does not already provide.

Multi-step retrosynthesis planning seeks to decompose a target molecule into commercially available building blocks through a sequence of feasible reactions. The vast combinatorial search space makes this task challenging even for expert chemists. Traditional methods combine tree search with offline-trained value networks that score candidates in isolation, without reasoning about complete multi-step routes. Recent work leverages Large Language Models (LLMs) for this task, but relies on simple interfaces that limit exploration of the full search space. We introduce RetroAgent, an LLM agent that bridges symbolic search and neural reasoning through a harness with structured memory. Through memory and chemistry tools, the agent observes the full search state, including explored routes, available alternatives, and properties of intermediates, enabling informed decisions grounded in both global progress and domain knowledge. Experiments on in-distribution and out-of-distribution benchmarks demonstrate that RetroAgent delivers strong performance and generalization.

This paper describes DS@GT ARC's third-place solution to the PlantCLEF 2026 challenge on multi-species plant identification in vegetation quadrat images, where systems must predict every species present in high-resolution (~3000 x 3000 pixel) plot photographs while training only on single-label images of individual plants. The pipeline is built around a fine-tuned DINOv2 ViT-L/14 classifier applied over a multi-scale tile decomposition of each quadrat, with per-tile predictions blended with a FAISS kNN retriever and post-processed by source-aware temporal fusion across repeated plot visits, a habitat-fit demotion that injects geographic and altitude priors from the training data, and a South-Western Europe geographic mask. Habitat-fit demotion and multi-scale aggregation are the largest individual contributors in the ablations. Two complementary training-centric directions, a cross-region transformer with noisy-student distillation on the LUCAS dataset and a label-as-query transformer decoder over synthetic CLS-domain pseudo-quadrats, yielded null results. An inference-time augmentation with instance-aware segmentation crops also did not improve performance. The selected submission reaches a private-leaderboard macro-F1 of 0.43902 (third place; public 0.51096); an unselected configuration of the same pipeline scored above 0.45 on the private set. Code: https://github.com/dsgt-arc/plantclef-2026.

While reinforcement learning with verifiable rewards (RLVR) is widely used to improve the reasoning capabilities of large language models (LLMs), the generalizability of the resulting models remains poorly understood. In this work, we establish the first non-vacuous generalization bounds for parameter-efficient RLVR fine-tuning at the billion-parameter scale. Our approach adapts PAC-Bayes compression bounds to this setting, and addresses the inherent stochasticity of token generation by applying the Gumbel-max reparameterization trick. To operationalize these bounds, we propose the Progressive RLVR framework, which integrates RLVR with on-policy distillation, TinyLoRA, and model quantization. Progressive RLVR empirically retains 84-97% performance of standard LoRA fine-tuning while producing models that are 14,796x more compressible. We show that this framework yields non-vacuous generalization bounds in four domains: mathematical problem-solving, programming, general-knowledge reasoning, and Text-to-SQL. Our bounds exceed the accuracy of the base model by 9-51% and lie within 6-11% of the accuracy of the fine-tuned models.

Complex knowledge base question answering (KBQA) is commonly approached through either information retrieval over a question-specific subgraph or semantic parsing into an executable logical form. We study the latter paradigm. Recent large language model agents make semantic parsing interactive: they alternate between reasoning, querying the knowledge base, and extending a partial SPARQL query. This interleaving reduces reliance on one-shot generation, but makes the quality of KB grounding depend on what the interaction tools expose. Existing agents retrieve or prune candidate properties mainly through lexical relevance and instance-level observations, without systematically conditioning on entity types, property domains and ranges, or the expected answer type. We call this failure mode type-blind grounding. It enlarges the grounding search space and often produces plausible-looking but semantically incompatible triple patterns that execute to empty results. We propose SAGA (\underline{S}chema-\underline{A}ware \underline{G}rounding for \underline{A}gentic Text-to-SPARQL Generation), a training-free framework that turns property exploration into a schema-constrained grounding operation. SAGA maintains a persistent bidirectional type state, filters known-incompatible property candidates at construction time, presents the remaining graph patterns in a compact schema-annotated format, and handles missing schema information permissively through empirical and trace-local evidence. Across nine benchmark settings over Wikidata and Freebase, SAGA achieves the highest F1 on all nine settings and the highest exact-match accuracy on eight, while reducing empty-result queries across all reported Wikidata settings.

Machine-learning force fields (MLFFs) are reliable only near their training distribution, making efficient construction of diverse training sets a major bottleneck for both train-from-scratch and foundation fine-tuning workflows. Active learning can reduce this cost, but standard model-committee uncertainty is impractical for foundation MLFFs because each committee member requires a separate fine-tuning run. We present an active-learning workflow based on last-layer-projection regression (LLPR), a forward-pass-cheap per-configuration uncertainty estimator. Across molecular, condensed-phase, and electrolyte systems, LLPR identifies compact, high-value training sets that recover full-data accuracy using only a small fraction of electronic-structure labels. In foundation-model fine-tuning, LLPR-selected configurations approach the practical fine-tuning ceiling with substantially fewer labels than random selection. In iterative electrolyte fine-tuning, LLPR detects unphysical local coordination before new labelling, calibrates uncertainty in force-error units, and supports retrospective convergence monitoring. The resulting models reproduce reference density and Li–O coordination structure, providing a scalable uncertainty-quantification strategy across MLFF training regimes.

Disordered metamaterials feature microstructures with inherent randomness and irregularity, enabling them to achieve broader property coverage and superior performance unavailable in their regular counterparts. Despite their promise, designing disordered microstructures is substantially harder than designing regular ones. Their design remains trapped between manual parameterizations with limited expressiveness, and generative AI that is data-hungry and struggles to generalize. To address these limitations, we propose a generative design framework based on Neural Cellular Automata that dynamically grows complex microstructures through learned local interaction rules, inspired by the self-organizing processes in natural materials. This framework requires only a single training template, yet accommodates diverse disordered microstructures and adapts to irregular domains and arbitrary discretizations. By manipulating the learned local rules, we can steer the growth process to generate microstructures unseen during training, providing control over orientation, anisotropy, and directional thickness without retraining. As a dynamic, local growth process, it naturally produces spatially varying microstructures that transition smoothly to enable location-specific mechanical properties. We demonstrate this in a multiscale mechanical cloaking design, where microstructures vary across the space to meet an optimized heterogeneous property distribution. Our design enables excellent cloaking performance without complicated post-processing and incompatible assembly common in existing methods. This data-efficient, generalizable approach opens access to previously intractable disordered materials for biomedical implants and soft robotics.

This paper details the DS@GT ARC team's approach to BirdCLEF+ 2026, multi-label detection of animal vocalizations in soundscapes from the Pantanal wetlands. The 2026 edition adds about an hour of labeled soundscapes, shifting the task toward supervised pipelines fit to the labeled set. First, we build a competitive supervised baseline that ensembles a frozen Perch v2 backbone, a trained HGNetV2-B0 sound-event-detection network, and a non-bird prototypical head, reaching a private leaderboard score of 0.936 at rank 1894 within a 90-minute CPU budget. Second, we ask whether token-based representations can compete, contrasting codec representations from neural audio codecs against semantic representations from foundational embeddings. We compare two bioacoustic specialist models against four token-based encoders trained on AudioSet. The repository for this work can be found at https://github.com/dsgt-arc/birdclef-2026.

Noise injection is a well-known technique in stochastic optimization. We report its surprising effectiveness with an interleaved (on-off-on-off...) rather than the usual monotonic decay schedule. We present a theoretical analysis of noise injection, which confirms that corruption by impulse noise approximates a Jacobian regularization, whereas Gaussian noise acts as a curvature penalty. This regularization behavior has been invoked to explain why noise injection increases model robustness. But the interleaved nature of our proposed schedule produces superior results even for the optimization objective: mixing phases of noisy data permits the optimizer to escape local minima and increase exploration without the risk of catastrophically forgetting the important features from the clean data. To stabilize this training scheme against the rapid changes of the loss when switching between clean and noisy data, we introduce a gradient-norm stabilization technique that scales noisy updates based on clean gradient magnitudes. We compare this method with other common augmentation methods and find substantial improvements in corruption tolerance and robustness to real-world distribution shifts on CIFAR-100-C, ImageNet-C, and ImageNet-R for ResNet and ViT architectures, with the best results being achieved by stacking our method on top of other augmentations. Through saliency and attention maps we show that the effect of interleaved noise injection stems from penalizing the failure modes encouraged by the inductive bias of the models: impulse noise works against the locality bias of convolutional (ResNet) architectures, and Gaussian noise reduces the tendency of attention-based models to pick up large-scale spurious features. Interleaved noise injection is therefore an effective tool to improve the test performance on clean, noisy, and out-of-distribution data at essentially zero computational cost.

We study Dynamical System Autoencoders (DSAE) for LiDAR point-cloud classification using spatial coordinates and Product Coefficient feature augmentations. The experiments compare separately trained DSAE architectures at encoder depths $K=1,\ldots,5$ and evaluate the resulting hidden representations with Random Forest, kNN, and a majority-class Dummy baseline. The main finding is a hidden-state collapse at $K=5$. For both xyz and xyz plus Product Coefficient inputs, the hidden-state standard deviation falls to the order of $10^{-5}$, while all three classifiers attain the same macro F1 score of $0.224688$. We prove that between-class hidden scatter is bounded by total hidden scatter, which in turn is controlled by the reported hidden-state variance. Thus a nearly constant hidden representation cannot retain substantial class-separating structure. Product Coefficients neither improve pre-collapse macro F1 nor prevent the $K=5$ collapse in the present DSAE setting. These results identify large-depth representation collapse as a concrete failure mode for DSAE LiDAR classification.

Generalist robot manipulation policies trained on large, diverse datasets have shown remarkable promise across a wide range of tasks. However, rigorously evaluating these policies remains a fundamental challenge. Real-world performance depends on a large combinatorial space of task factors including object poses and camera viewpoints, making full, exhaustive evaluation intractable. Additionally, real hardware evaluation is slow and resource-intensive, so current practice is to use narrow test suites that can miss critical failure modes and misrepresent true deployment readiness. We propose an active evaluation framework that addresses this challenge by treating policy evaluation as a sequential experimental design problem. Our approach fits a probabilistic surrogate model over a structured space of task factors and adaptively selects evaluation configurations to maximize information gain over the policy's performance distribution, allowing for sample-efficient characterization of policy behavior across unseen conditions and a systematic identification of failure-prone regions. We conduct 2331 real-world evaluations across 3 tasks with 3 factor variations and find that our approach typically saves the evaluator at least 20-40% of trials compared to typical random testing.

Mechanistic studies of large language models localize function in parameters, attention heads or neurons, but those parts do not specify how one inference unfolds. We treat the ordered layer-state trajectory as the empirical process and test whether layer order defines a measurable computation. Across representative Qwen, Llama and Gemma checkpoints, realized order produced paths 1.60–2.24 times as aligned with their endpoint chord as endpoint-preserving shuffles. The order effect transferred to a lexical task, but was stronger in three random initializations of the same Qwen architecture, identifying an architecture-compatible scaffold rather than a training-specific law. Direct hidden states provided stronger mechanism coordinates than vocabulary readbacks, and a task-defined coordinate made candidate competition computable. After additive interventions failed, local transition actuators selectively controlled internal trajectories across checkpoints. A confidence gate then drove 17 of 87 held-out Qwen closure trajectories across the full-vocabulary top-1 candidate boundary while leaving all 203 non-closure top-1 outputs unchanged. Safety-matched action reassignment performed similarly, locating the gain in selective gating and the bounded actuator family rather than fine-grained dose assignment. Layer order exposes an observable, computable and behaviourally consequential process of transformer inference without establishing a complete theory of LLM dynamics or control of answer correctness.

Diffusion Large Language Models (dLLMs) have recently emerged as a promising alternative to conventional Auto-Regressive (AR) Large Language Models (LLMs). By leveraging bidirectional attention and parallel decoding, dLLMs enable more efficient generation. However, they require a carefully designed denoising scheduler at inference time (absent during training) whose choice significantly impacts generation quality. While confidence-based heuristic schedulers have shown strong empirical performance, they suffer from two critical failure modes: EOS Overflow and Proximal Bias. Through in-depth analysis of the Transformer's attention patterns, we reveal that these failures stem from certain positions assigning disproportionately high attention weights to invalid tokens (e.g., [MASK] and [EOS]), which produce misleading confidence signals. Building on this insight, empirical evidence shows that valid attention scores can provide complementary guidance to conventional confidence-based heuristics, yet no single metric consistently excels across all scenarios, implying that the optimal denoising trajectory is highly context-dependent. To address this problem, we propose a lightweight evolutionary heuristic scheduler optimized using the Covariance Matrix Adaptation Evolution Strategy (CMA-ES). Our scheduler dynamically integrates multiple heuristic features with a contextual mean-field embedding, while requiring only 393 trainable parameters. Evaluated on LLaDA and Dream across four reasoning and planning benchmarks, our method consistently outperforms strong baselines, including conventional heuristics, block auto-regressive methods, and recent State-Of-The-Art (SOTA) approaches. To the best of our knowledge, it represents the most parameter-efficient neural scheduler to date. Our code is available at https://github.com/RS2002/Evo-Denoise .

A robot's learned model of its own dynamics is only valid until those dynamics change: actuators wear, payloads shift, and joints stiffen. A model-based agent that keeps training as if nothing happened adapts slowly, dragged back by a replay buffer full of stale experience. We present Changepoint-Aware World Models (CAWM), a DreamerV3 agent that detects an abrupt dynamics shift from its own internal prediction error, using an online CUSUM test against a rolling baseline that fires only on abrupt change rather than on slow learning drift. It then forgets stale replay, keeping the learned representation while flushing obsolete data. On simulated locomotion under two robot-relevant shifts, doubled gravity and halved actuator gain, CAWM recovers substantially faster than passive retraining. It also beats a strong baseline that respawns a fresh dynamics model on detection, the deep-world-model analogue of model-bank methods. With the response triggered at the shift, CAWM gains +95 to +153 return in the first 30k post-shift frames over three seeds, while matching that respawn at asymptote. Running the detector in closed loop reproduces this gain on the gravity shift. The benefit holds across both shift types, and is largest when the shift is severe enough that old data is genuinely obsolete.

This paper investigates collaborative speculative decoding (CoSD), a distributed large language model (LLM) inference framework in which an on-device small language model (SLM) drafts candidate tokens and a server LLM verifies them. Existing CoSD methods assume a shared vocabulary between the SLM and the LLM and incur substantial communication load because residual resampling requires token distribution exchange between the user device and the edge server. To address these limitations, we propose cross-vocabulary CoSD (X-CoSD), a lossless and communication-efficient CoSD framework for heterogeneous SLM-LLM vocabularies. X-CoSD is built on hybrid resampling (HR), which splits residual resampling across the common-vocabulary region on the device and the LLM-only region on the server, so that distribution transmission is required only for the common-vocabulary region. We further propose X-CoSD-E, an enhanced variant based on server resampling with device verification (SR-DV), in which the server sends only replacement candidates sampled from the server LLM and their corresponding probabilities for local verification at the device. We prove that both X-CoSD and X-CoSD-E preserve the server LLM distribution, and experiments show that they significantly improve token generation speed while maintaining generation quality comparable to that of the server LLM.

Mobile and wearable sensing enables longitudinal observation of behavior, yet translating these signals into meaningful mental health constructs remains difficult. We introduce a clinician-in-the-loop benchmark for evaluating whether large language models (LLMs) can generate evidence-grounded Brief Hierarchical Taxonomy of Psychopathology (B-HiTOP) item profiles from passive sensing, ecological momentary assessment (EMA), and questionnaire evidence. Using the Generalization of Longitudinal Behavior Modeling (GLOBEM) dataset, we construct 14,592 participant-day instances and align multimodal evidence to 29 B-HiTOP items across five spectra. Since GLOBEM lacks B-HiTOP responses, we evaluate evidence compatibility (C) rather than diagnostic accuracy, separating substantive predictions from abstentions when evidence is insufficient for item-level scoring. Two-stage prediction improves C for EMA and questionnaire evidence, but reduces C under passive sensing and combined evidence and produces more conservative score distributions across models, spectra, and evidence settings. Overall, semantic abstraction helps organize heterogeneous self-report evidence while becoming an information bottleneck for indirect behavioral sensing signals.

Road safety monitoring has historically been reactive, relying on crash-record analysis after fatalities and injuries have already occurred. Proactive identification of high-risk locations and dangerous driving behaviour before incidents occur is a critical but underexplored challenge. This paper addresses this gap using connected vehicle telemetry data from Greater Sydney, Australia, to detect and forecast near-miss risky driving events at the Local Government Area (LGA) level. Risky driving is quantified through g-force thresholds (hard braking >0.6g, harsh cornering >0.47g, harsh acceleration >0.5g), and spatio-temporal heatmaps are constructed to identify high-risk zones. Eight predictive models are benchmarked across three families: ensemble learning (Random Forests, XGBoost, LightGBM), deep learning (LSTM, N-BEATS), and classical time-series methods (ARIMA, Exponential Smoothing, Prophet). ARIMA achieves the lowest mean absolute error (MAE: 162.21), performing comparably to LSTM (MAE: 163.92) and outperforming all ensemble methods, with N-BEATS reaching an MAE of 180.75. These results demonstrate that parsimonious time-series models are competitive with deep learning approaches when training data volume is limited. The study highlights the potential of IoT-based connected vehicle data to support proactive road safety interventions, with Sydney's inner and western LGAs (CBD, Parramatta, Bankstown) identified as persistent high-risk zones warranting targeted policy action.

The performance of deep learning models at scale relies heavily on how effectively high-level mathematical operations are mapped to underlying physical hardware. While high-level tensor frameworks provide flexible abstractions, their execution models inherently lack the whole-graph visibility required to maximize hardware utilization, often forcing a reliance on opaque, hand-written kernel libraries for complex operations like Attention. To bridge this gap, we present the next iteration of Nova, an automated end-to-end JIT compiler that achieves absolute control over hardware mapping by synthesizing fine-grained kernels directly from the computation's structure. In this work, we extend Nova's compilation pipeline to natively support full Transformer architectures. By capturing eager executions and unifying forward and backward passes into a single value-semantic dialect, Nova unlocks aggressive whole-graph optimizations. Rather than relying on rigid, pre-compiled library calls, Nova focuses on extensive cross-operator fusions, collapsing complex causal attention sub-graphs, element-wise operations, and memory-bound normalizations directly into single fused kernels to drastically reduce global memory roundtrips. In our evaluations training a full GPT-2 architecture on Ada 6000 GPUs, Nova demonstrates superior end-to-end throughput, averaging 441K tokens/second compared to 406K for our own eager execution and 405K for torch.compile. By drastically reducing memory-bound overheads through compiler-native fusion, Nova enables efficient full LLM compilation on modern hardware while strictly maintaining numerical parity.

The lottery ticket hypothesis proposes that large random neural networks contain sparse subnetworks that can match the performance of dense models after comparable training. A stronger version asserts that sufficiently overparameterized random networks contain subnetworks that are already accurate before any weight training. Existing theory establishes that such strong lottery tickets exist, but reliable extraction remains difficult. We revisit edge-popup, a frozen-weight score-training method for extracting strong tickets, and identify layerwise sparsity selection as a central bottleneck. We introduce double-scoring, an augmented score-space parameterization that replaces a layerwise sparsity search with optimization over enlarged score tensors. We prove that fixed-density masking in an augmented score space preserves access to all original-coordinate masks, and we show that the resulting method can be interpreted as edge-popup on a zero-augmented network. In controlled experiments, double-scoring substantially improves strong-ticket extraction over fixed-density edge-popup and pruning-at-initialization baselines, improves on the performance of rewound sparse-training topologies, and exhibits markedly lower sensitivity to sparsity hyperparameters. Ablations show that the gain is not merely due to additional trainable score parameters, but is tied to the augmented score-space competition that induces the effective original sparsity.

Reliable and low-latency NR-V2X communications are essential for smart mobility in dense urban environments. However, limited Road-Side Unit (RSU) density, frequent non-line-of-sight conditions, and highly dynamic vehicular topologies often prevent many Connected and Automated Vehicles (CAVs) from maintaining stable single-hop connectivity. Although multi-hop relay-assisted communication can extend infrastructure coverage, selecting relay links in real time under practical flow, capacity, and connectivity constraints remains challenging. Mixed-Integer Linear Programming (MILP) yields optimal multi-hop relay decisions, but its computational complexity scales sharply with network density, limiting real-time applicability. To address this, we propose a Learning-to-Optimise (L2O) framework based on Graph Neural Networks (GNNs) for real-time NR-V2X relay selection. Vehicular communication states are modeled as attributed graphs, where CAVs and RSUs are nodes and candidate radio links are enriched with propagation-aware features. An offline MILP oracle provides optimal supervision, while an edge-aware Graph Isomorphism Network (GINE) approximates oracle decisions with near-constant inference latency. Experiments on large-scale urban datasets generated by an integrated SUMO–GEMV2 simulation pipeline show that the proposed approach achieves connectivity comparable to that of the MILP oracle while reducing execution time by orders of magnitude. The framework enables cost-effective enhancement of urban V2X connectivity by leveraging existing vehicular assets and supporting scalable, real-time NR-V2X operation in smart city environments.

Grokking – the delayed generalization of neural networks long after they have memorized their training data – wastes thousands of training epochs and is notoriously unpredictable. Building on the recent result that Transformer attention is formally isomorphic to a thermodynamic system, we treat the variance of attention logits as a specific heat Cv and show that its peak reliably precedes the generalization transition. We introduce CvAdamW, a drop-in AdamW variant that monitors Cv online and injects thermal energy by dynamically scaling weight decay when a phase transition is detected. Through a strictly iterative development process we identify three failure modes – initialization noise, mini-batch micro-ripples, and slingshot blinding – and resolve them with a memorization gate and an exponential-moving-average shock absorber. On modular arithmetic (a+b mod 97), CvAdamW enables grokking at epoch 2802 in a 4000-epoch budget where the baseline never groks. We further propose a scale-invariant z-score reformulation that removes task-specific hyperparameters, and evaluate it across 10 paired seeds. A paired analysis shows the cold-start variant reduces mean grokking latency by 257 epochs (6.0%; median 166 epochs; Wilcoxon p=0.049, Cohen's d=0.68, bootstrap 95% CI [53,489]), improving 8 of 10 seeds; on this single task Cv peaks before grokking in all 10 seeds. Our results indicate that neural networks may expose detectable precursors of impending generalization transitions, and that a physically motivated, proportional intervention can facilitate generalization within a fixed compute budget. Code and data are public.

Effective molecular representation learning is crucial for accurate molecular property prediction. Recently, numerous self-supervised learning (SSL) approaches leveraging 3D GNNs have been developed to capture comprehensive 3D structural information for drug discovery. However, existing methods lack explicit physical constraints and are highly susceptible to geometric noise induced by coarse empirical force fields during large-scale pre-training.Furthermore, they overlook dynamic feature modulation during downstream adaptation, often resulting in catastrophic forgetting and negative transfer. To address these limitations, we introduce SenCos-GEM, a novel explicitly decoupled geometry-enhanced molecular representation learning framework that incorporates SENet-calibrated and law-of-cosines-constrained enhancements. SenCos-GEM employs a physics-guided geometric consistency loss based on the law of cosines to derive high-fidelity and mathematically invariant 3D spatial priors. In addition, lightweight Squeeze-and-Excitation (SE) modules are integrated into the backbone as task-specific adapters, while a dual-modulation prediction head combines Feature-wise Linear Modulation (FiLM) and SENet mechanisms to enable dynamic feature recalibration. SenCos-GEM demonstrates highly competitive performance across diverse classification and regression tasks on MoleculeNet benchmark, establishing new state-of-the-art results specifically on 3D conformation-sensitive regression tasks, such as FreeSolv, Lipophilicity, and QM9, achieving relative error reductions of 12.9% (RMSE), 5.3% (RMSE), and 8.2% (MAE), respectively. Moreover, our model exhibits superior capability in distinguishing stereoisomers and discriminating conformational perturbations, underscoring its robust spatial modeling performance. Collectively, SenCos-GEM represents a significant breakthrough in accurate molecular property prediction.

Lipschitz continuity is a fundamental property of neural networks that characterizes their sensitivity to input perturbations. It plays a pivotal role in deep learning, governing robustness, generalization and optimization dynamics. Despite its importance, research on Lipschitz continuity is scattered across various domains, lacking a unified perspective. This paper addresses this gap by providing a systematic review of Lipschitz continuity in deep learning. We explore its theoretical foundations, estimation methods, regularization approaches, and certifiable robustness. By reviewing existing research through the lens of Lipschitz continuity, this survey serves as a comprehensive reference for researchers and practitioners seeking a deeper understanding of Lipschitz continuity and its implications in deep learning.